The function of the control arms is to transfer movements applied to them by an operator into control instructions for an instrument or a system, usually a remote robot called a slave arm or a computer simulation. When the control arm is provided with a sufficient number of degrees of freedom, the operator can control it in translation and in rotation in space.
There is a wide variety of forms of anus used in robotics. In the most traditional form, they are composed of a chain of segments connected to each other by articulations or sometimes by other types of joints; this arrangement is called “in series”. But there are disadvantages with this type of arm whenever there are a large number of segments.
Thus, there is play at joint mechanisms which accumulate and eventually produce a significant imprecision on the position of the free end of the arm. Furthermore, the motors that normally need to be added to the arm to control the states of the joints in order to modify their configuration or to hold them in a fixed position regardless of the external forces applied to them, and which are frequently the heaviest part of the arms, cause excessive bending moments that can make it necessary to reinforce the structure of the segments and therefore further increase the weight of the arm, making it inconvenient to manipulate. It was proposed to displace the motors onto the fixed base on which the arm is mounted, but this solution would require installing transmissions between the motors and the joints controlled by them, which is not always possible and in any case contributes to making the arm complex.
This is why in the more recent history of robotics, designers have been interested in constructions using several (two or more) arms in parallel, in which the distal ends are connected together. For a similar number of degrees of freedom, the branches of arms built in this way are less complex than in robots in series, which significantly attenuates the disadvantages of the imprecision in the position of the arm and the weight of the branches. However, there are specific limitations with this type of robot. They may thus be difficult to control to a given state, due to the fact that their kinetic complexity is greater; their working range is usually smaller than the working range of arms in series, since it is limited by the working range of different branches in parallel and by collisions between segments of the different branches; finally, a fault correlated to the previous fault is that the number of singularities, which are configurations that should be avoided since the robot may be subjected to uncontrolled movements at these points, is generally greater.
Singularities correspond to local disappearances of degrees of freedom or uncontrolled movements. Singularities corresponding to uncontrolled movements are specific to parallel robots, but all singularities restrict the usage range of the arm. This disadvantage is more pronounced with master arms, which are not designed to apply repetitive movements or movements known in advance and that are controlled by hand without thinking about singularities, and which therefore may occur unexpectedly. This is why they must be limited, while transferring them to the ends of the working range. It is concluded from these various remarks that arms in parallel are attractive as master arms due to the convenience in manipulating them, but in reality their specific defects make most of them inappropriate for this application. The invention deals with a particular arrangement of arms in parallel, for which the main advantages are a significant reduction in singularities and considerable ease of control, which means that the movement required to reach a required state can be imposed without difficulty.